In order to build a useful and comfortable in-car human machine interface systems, the information presentation method should be easy to understand (low mental workload) and one should be able to respond with ease to the information presented (low response workload). We are making efforts to establish an evaluation method that would differentiate between mental workload and response workload. Here, we present the results of our trial using brain waves measurements (Eye Fixation Related Potentials). We focus on the relation between P3 latencies and drivers response workload compared to mental workload in a task involving eye movements. Previous experiments showed that P3 latency correlates strongly with the amount of information presented. The current experiment shows that P3 latencies seem to be independent to the type of response the subject is requested to perform. When asked to push a button with their right hand after detecting a Landolt ring pointing right, subjects react faster than when pressing the button with their right hand in response to a ring pointing left. The same situation is replicated whether the subjects is presented the visual stimulus in the center of their visual field, or a driving scene video is played in the center and the target stimulus is presented in a lower left position similar to in car navigation systems display position. In both situations P3 component latencies are similar. While needing to expand the experiment to a larger number of subjects, and to improve the measurements method for higher stability, this experiment combined with our previous findings show that P3 latencies seem to be a good independent measure for cognitive workload.